uTracer (Micro Curve Tracer)

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

Post Reply
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

I'd be happy to trace a 6GH8 as a 7199 surrogate in triode and pentode connection. Just PM me.

Rog , I played with an earlier version of Nick B's GUI when I was looking for a way to run the uTracer using a Mac. It does have some nice features, but I've solved that problem by running Windows 8 using Bootcamp. If I want to make a really nice plot I just export the data file to Excel.
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Another mod for the uTracer has been suggested by its designer R Dekker, which is doubling up the inductors on the anode and screen supply boost converters. This increases the reservoir capacitor charging rate and speeds up the process of tracing a set of curves. The "Quick Test" is noticeably faster too. These are surface mount parts, so a second inductor for each high-voltage supply can be attached on the underside of the PCB, in parallel with the originals.

While I was at it, I added another switch to the front panel (it's the lower one on the left in the dramatic action shot below). It connects a pair of 10k 1% resistors from the anode and screen terminals to the cathode terminal. With this switch engaged, a calibration test set-up can be recalled and run using known resistances to check the anode and screen current measurement accuracy. This enables a quick check and adjustment of the current amplifier calibrations.

This calibration switch integrates nicely with the "continuous mode" switch just above it, which directly connects the anode and screen reservoir capacitors to the panel terminals. Selecting continuous mode permits a check of the boost converter voltage calibrations from the front panel using a DVM. The switch connections are arranged such that the 10k current amplifier calibration resistors cannot be connected when continuous mode is selected. They would be toasted in short order if that happened and a test sequence were to be started.
You do not have the required permissions to view the files attached to this post.
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

More stupid uTracer tricks...

Here is a trace of a simple neon voltage tester of the type used to check for "live" mains circuits. It consists of a neon bulb in series with a current limiting resistor (see photo). The I-V trace has been done in continuous mode to get a sharp rise in current at the starting voltage and a crisp transition to the resistive load line. Extending the resistive load line indicates that the voltage drop after current begins to flow is reduced to ~55V from a starting voltage of ~72V, and the slope of the line suggests that the current limiting resistor is a 220k.

Why do this? Because a voltage regulator tube (0A3, e.g., of which I don't have an example) can be tested using the uTracer in the same way.

An 0A3 data sheet indicates a starting voltage of 100V, and an operating voltage of 75V at 5 to 40 mA. To reach 5 ma at the starting voltage a 5k current limiting resistor would be required (25V/0.005A), and limiting the current to 40 mA means the maximum allowable anode voltage would be 200V (5k ohms*0.04A), plus the 75V operating voltage drop, or 275V. Limiting the trace to 200V would be safe and result in a nominal maximum current of (200V-75V)/5k = 25 mA. This is more than the uTracer can supply in continuous mode, so a pulsed measurement would be necessary. If the regulator tube could sustain a lower current flow, a larger resistor, 10k say, might be used to find the starting voltage in continuous mode.
You do not have the required permissions to view the files attached to this post.
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Shown below is a jury-rigged test of one of the triode sections in a GE 6C10 Compactron, one of a family with 12-pin bases that appeared near the end of the tube era. I only have 7, 8, and 9-pin sockets hard-wired in, but any tube type can be connected and tested in this way. If you look closely you can see the two hanging loops of the series-connected filaments glowing below the bottom mica.

The socket array has ferrite beads on the wiring to prevent oscillations, but those are bypassed when connecting directly to the ASGK jacks as shown. No oscillation issues were evident here, so it seems that the ferrites on the uTracer's PCB are sufficient, at least in this case, to keep things under control.

This particular tube type was used in Fender's Super Champ (1982-1985), where its three 12AX7-equivalent sections were used as the reverb recovery stage, a gain stage following mixing of the wet and dry signals, and as it's cathodyne phase inverter.

The .pdf below is a nice summary of Compactron types and the conventional equivalents to their internal sections.
You do not have the required permissions to view the files attached to this post.
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Zero grid bias anomaly explained!

Some time back I compared Ia-Va traces at 0V Vg for a 12AX7 where the grid voltage was first provided by the uTracer and then fixed at zero by tying the grid to the cathode: http://ampgarage.com/forum/viewtopic.ph ... 152#329152 The traces didn't agree very closely, and the reason remained a mystery. It is known that the uTracer can't really set Vg at 0V; the upper limit is actually about -50mV due to the saturation voltage of the grid supply output transistor. That isn’t enough to explain the difference in the traces, so the observed behavior still suggested that the voltage was not being set accurately as it approached zero.

Ronald Dekker looked into this and found that the grid bias was in fact falling below the set voltage during the 1 ms measurement pulse. He found the cause to be a rise in the ground reference for the PIC’s PWM signal due to heavy current draw from the LED’s in the anode and screen voltage optocouplers. Although this problem existed in the uTracer 3 it became much more noticeable in the uTracer 3+ (400V upgrade), where the optocoupler current is almost doubled.

The grid bias voltage is produced from a PWM signal (5V, 20 kHz, variable-duty-cycle square wave), which is converted to a 0-5V variable DC voltage by passing it through a low-pass filter. This DC voltage is then amplified -10x to make a 0 to -50V source. The rise in the PWM voltage is small (10-20mV), but the grid bias voltage is then shifted downward by ten times as much (100-200mV). This Vg shift is enough to affect the accuracy of the Quick Test for tubes with low grid bias voltage like the 12AX7, where the specified -2V becomes -2.1 to -2.2V, producing a low anode current reading.

Once again, Ronald came up with a simple fix that solves the problem without requiring any modification to the PCB: Lower the corner frequency of the filter by increasing the values of its two capacitors (C10 and C11). This has the side effect of delaying the grid bias circuit’s response to the rise in the PWM reference voltage until after the measurement pulse is finished. Brilliant!

In the plot below a 12AX7 is traced with 0.1V increments from Vg = 0.5V to Vg = 0V. A second Vg = 0 line is traced with the grid tied to the cathode (upper-most black line). Capacitors C10 and C11 have been increased to 220nF and 330nF respectively. There is now a much smaller difference between the two Vg = 0 lines, and even better, if the blue Vg = 0 curve is really at -50mV, it looks to be in just about the right place, halfway between the black curve (true Vg = 0) and the -0.1V green curve.
You do not have the required permissions to view the files attached to this post.
John_P_WI
Posts: 1456
Joined: Wed Aug 15, 2007 4:29 pm
Location: Wisconsin

Re: uTracer (Micro Curve Tracer)

Post by John_P_WI »

Hi Martin,

I just received my kit today. Looking forward to the build and challenges ahead.

I'll have to read back through this entire thread for ideas and inspiration.

Thanks!
John_P_WI
Posts: 1456
Joined: Wed Aug 15, 2007 4:29 pm
Location: Wisconsin

Re: uTracer (Micro Curve Tracer)

Post by John_P_WI »

Dbl
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

That's great, John, I'm sure you'll find it an interesting and useful project. After almost three years of use I'm still very happy with it.

Here's a picture of an indicator tube under test. There were commonly seen in radios and test equipment from the 1930's until the end of the tube era. I saw this NIB JAN tube offered for a good price, and picked it up just for fun. The uTracer is operated in continuous mode for this, with the screen supply connected directly to the "Target" (the part that glows), and the anode supply connected to the anode through a 1M load resistor. The size of the gap in the "display" pattern is controlled by the grid voltage. The operation of the tube can easily be observed in pulsed mode too. I find that the the flexibility of a patch cord set-up is ideal for cases like this.
You do not have the required permissions to view the files attached to this post.
User avatar
JazzGuitarGimp
Posts: 2355
Joined: Mon Jul 23, 2012 4:54 pm
Location: Northern CA

Re: uTracer (Micro Curve Tracer)

Post by JazzGuitarGimp »

Ah-ha! The old "Magic Eye Tube" trick!
Lou Rossi Designs
Printed Circuit Design & Layout,
and Schematic Capture
John_P_WI
Posts: 1456
Joined: Wed Aug 15, 2007 4:29 pm
Location: Wisconsin

Re: uTracer (Micro Curve Tracer)

Post by John_P_WI »

Hi Martin,

Yes, I love the magic eye tubes, they are very cool. I also like nixie tubes that unfortunately have no audio value...

Any suggested mods right off the bat would be appreciated like an external heater supply switch etc. I think for the time being I'll only wire an octal and two 9 pin sockets - one dedicated for testing both halves of a 12A_7 simultaneously. I'll also include jacks for heater current and voltage etc.

Honestly I'm most worried about getting the communication working with my old XP notebook.

Thanks again.

John
Matt D
Posts: 170
Joined: Fri Apr 22, 2011 6:21 pm

Re: uTracer (Micro Curve Tracer)

Post by Matt D »

I have a question about this tester.

I've noticed when testing power tubes and higher current preamp tubes like 12AT7's, 12AU7's, and 12BH7's that the tube needs to be flowing current for some time before it "settles in" to a stable plate current. For preamp tubes like those mentioned above about 1 1/2 to 2 minutes seems sufficient. For power tubes 5-10 minutes of current flow works depending on the tube. I've tested hundreds of preamp tubes like this using an Amplitrex in fixed-bias mode and using bench supplies to run power tubes at specified plate and bias voltages while measuring the plate current.

If the HT in this tester comes on in an instantaneous pulse, how accurate can the measurement be for tubes that flow 10mA of of plate current and higher? For lower current tubes like the 12AX7 maybe this wouldn't be as much of an issue. Also, I might have misread the information on the website. Is current flowing in multiple pulses over a period of time or does the Micro-Tracer use a single, instantaneous pulse?
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

John_P_WI wrote:Any suggested mods right off the bat would be appreciated like an external heater supply switch etc.
From what I can see in the current assembly manual the latest kit versions include the board-level mods described above, except for the doubled-up inductors. Beyond that, I exclusively use a bench supply for the DUT's heater. See here for the explanation: http://ampgarage.com/forum/viewtopic.ph ... 841#295841 The schematic representation of the heater switch is shown below, along with the mods for the continuous mode and current amp calibration switches. Note I did not use the supplied ferrites in the heater connections to minimize lead inductance, and that the heater terminal block, D19, and D20 are mounted on the board.
John_P_WI wrote:Honestly I'm most worried about getting the communication working with my old XP notebook.
I've been using a USB TTL adapter as described here: http://ampgarage.com/forum/viewtopic.ph ... 764#329764
You do not have the required permissions to view the files attached to this post.
User avatar
martin manning
Posts: 13079
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Matt D wrote:I have a question about this tester.

I've noticed when testing power tubes and higher current preamp tubes like 12AT7's, 12AU7's, and 12BH7's that the tube needs to be flowing current for some time before it "settles in" to a stable plate current. For preamp tubes like those mentioned above about 1 1/2 to 2 minutes seems sufficient. For power tubes 5-10 minutes of current flow works depending on the tube. I've tested hundreds of preamp tubes like this using an Amplitrex in fixed-bias mode and using bench supplies to run power tubes at specified plate and bias voltages while measuring the plate current.

If the HT in this tester comes on in an instantaneous pulse, how accurate can the measurement be for tubes that flow 10mA of of plate current and higher? For lower current tubes like the 12AX7 maybe this wouldn't be as much of an issue. Also, I might have misread the information on the website. Is current flowing in multiple pulses over a period of time or does the Micro-Tracer use a single, instantaneous pulse?
Anode and screen voltages are applied in 1 ms pulses. For each data point, a number of measurements (1 to 16) are made and averaged. The number of averages used is determined automatically by the firmware, or it can be set manually by the user. Averaging is used to reduce noise in the measurement, and the pulsed measurement technique allows measurements at points well beyond normal steady-state operating limits.

1 ms is plenty of time for the current flow at the test voltages to stabilize, but the thermal state of the various elements inside the tube will affect their dimensions and spatial relationships, and that could alter the electrical behavior of the tube. The thermal state of the instrument itself can alter the results too, as I found above by measuring fixed resistors over a 50 minute period, where ~1% drift was observed.

For a tube in a given ambient temperature environment, the thermal state is determined by the heat dissipated by the filament as well as the anode and screen dissipation, if current is flowing. The standard test procedure, though, is to apply power to the filament, allow the temperature to stabilize with no anode or screen current flowing, and then apply voltages for a short interval of time to make the measurement. In older testers, the test interval was the momentary push of a button. Cathode temperature is the dominant factor in emission, and filament current is the largest driver of that. In circuit, the tube's internal temperature will be higher due to the quiescent current flow, and higher still under signal conditions for Class AB applications.

After the heater has stabilized, how much difference do you observe in the measured characteristics (Ia, gm, ra, mu) after anode current is allowed to flow for a few minutes as opposed to testing with current flowing momentarily?
Matt D
Posts: 170
Joined: Fri Apr 22, 2011 6:21 pm

Re: uTracer (Micro Curve Tracer)

Post by Matt D »

Hi Martin, Thanks for your reply. In general I notice a 1.5 mA to 3 mA difference in plate current after letting a tube like a 12AU7 or 12AT7 flow current continuously for a couple of minutes. This does not happen with lower current preamp tubes like a 12AX7. Tomorrow I'll be around the Amplitrex and can give you some data from the same tube tested different ways. I'm thinking:

1. Heaters on for one minute, activate the B+ and measure immediately.
2. Heaters on for 3 minutes, activate B+ and measure immediately.
3. Heaters on for 1 minute, then flow current for 2 minutes and measure.
John_P_WI
Posts: 1456
Joined: Wed Aug 15, 2007 4:29 pm
Location: Wisconsin

Re: uTracer (Micro Curve Tracer)

Post by John_P_WI »

Thanks Martin for your reply and schematic. I'm sure I'll have many more questions once I get into it.

John
Post Reply